Why did the Concorde crash?
On July 25, 2000, Air France Flight 4590, a Concorde registered F-BTSC, was taking off from runway 26 right at Paris–Charles de Gaulle when one of its tires ran over a strip of metal lost by an aircraft that had departed five minutes earlier. The tire was destroyed. A large piece of the left wing’s fuel tank 5 was ripped out, and the leaking fuel ignited. Two left-side engines lost thrust, the landing gear would not retract, and the aircraft crashed onto a hotel in Gonesse. All 109 people aboard and four people on the ground died. BEA final report, synopsis and §1.1–1.2, pp. 14–18.
The large section of tank found on the runway bore no signs of impact. To explain how a burst tire could tear it out, the French investigation agency, the BEA, examined how the impact travelled through the full tank. The agency concluded that a piece of tire struck elsewhere and the transmitted force tore that section out from the inside. BEA findings and probable causes, pp. 174–176.
Watch the Concorde investigation
The companion episode follows the strip, the decision speed and the tank rupture in sequence, using simplified technical animations and clearly illustrative scenes.
Watch “The Crash That Killed the Supersonic Age” on YouTube.
A strip of titanium on the runway
The strip was 435 millimetres long, 29 to 34 millimetres wide and about 1.4 millimetres thick, cut from a titanium alloy. It came from a thrust-reverser cowl door on engine 3 of a Continental Airlines DC-10, flight COA 55 to Newark, which had taken off five minutes before Concorde. The BEA attributed its loss to maintenance that did not follow accepted practice. BEA §1.16.6, pp. 103 and 107; findings, p. 174; recommendation 4.3.4, p. 180.
The strip lay about 1,740 metres down the runway. Concorde’s tire No. 2, the front right tire of the left main landing gear, ran over it at about 175 knots, some 1,720 metres from the threshold. The crew had already passed the takeoff decision speed, V1, of 150 knots. BEA §1.1, p. 17; §1.12.1, pp. 59–61; §2, p. 161.
The destroyed tire threw debris at the underside of the wing. One piece of tire recovered from the runway weighed about 4.5 kilograms. The investigators calculated that a piece of that size striking at around 140 metres per second could supply the energy needed for the rupture. BEA §1.12.1.2, p. 60; §1.16.7, p. 115.
How did the fuel tank tear open?
Tank 5 was practically full, at about 94 percent of its volume. Takeoff acceleration pushed the fuel’s free surface to the front of the tank. Fuel barely compresses: when the impact pushed in the tank wall, displaced fuel pushed outward on neighbouring parts of the structure. The BEA called this an “indirect mode” of rupture. BEA §1.16.7.2–1.16.7.3, pp. 113–118.
Several pieces of evidence support that account:
- A large part of tank 5’s underside, found on the runway, bore no signs of impact and had been ripped away from the inside towards the outside.
- Another part of the tank’s underside, found at the accident site, had a puncture about 10 by 40 millimetres. The tank was penetrated there, but that puncture does not explain the large missing section.
- Firing tests with tire pieces and computer modelling confirmed that the indirect mode existed, although the equipment could not reproduce the exact rupture.
The report’s probable causes describe the rupture as a complex transmission of energy from an impact “at another point on the tank,” combining deformation of the tank skin and movement of the fuel. A hydrodynamic pressure surge, the shock studied in military projectile research, was considered only a possible contributing effect. The report’s findings state that a surge from a penetrating projectile could not have caused the ripping out of the piece found on the runway. BEA §3.1–3.2, pp. 174–176.
The fire and the final minute
A large fire, fuelled by the leak, broke out almost immediately under the left wing. The BEA identified two possible ignition sources: an electric arc in the landing gear bay, or contact with hot engine parts. It did not settle on one. At 14:43:13, the controller told the crew they had flames behind them. BEA synopsis, p. 14; ATC communications, p. 39; probable causes, p. 176.
The aircraft was past V1 and close to rotation speed. The BEA’s simulation showed that an attempt to stop would have led to a high-speed runway excursion. Engine 2 lost thrust and was shut down after a fire alarm; engine 1 also surged. The crew could not retract the landing gear, which probably helped hold the flame in place. BEA §2, pp. 165–166; probable causes, p. 176.
For about a minute, Concorde flew at around 200 knots and a radio altitude of about 200 feet, unable to gain height or speed. The crew said they were trying for Le Bourget. Engine 1 then lost thrust, and the bank to the left increased to 113 degrees. The aircraft crashed less than a minute and a half after the tire was destroyed. BEA synopsis, p. 14; §1.1, pp. 17–18; §2.1.4, p. 165.
What changed after the Concorde crash
Tire damage had harmed Concorde’s tanks before. On June 14, 1979, at Washington Dulles, burst tires caused small punctures in tanks 2, 5 and 6. BEA §1.16.4.2.1, p. 94. After Flight 4590, the BEA and the UK’s AAIB recommended suspending Concorde’s certificates of airworthiness, and the French and British authorities did so. BEA §4.1, p. 177.
Before Concorde returned to service, the airworthiness authorities required:
- flexible linings in tanks 1, 4, 5, 6, 7 and 8;
- reinforced electrical harnesses in the main landing gear bays;
- Michelin NZG tires and a modified anti-skid computer;
- operating changes, including inhibiting power to the brake ventilators during critical phases of flight.
The BEA also recommended changing certification requirements to account for tank damage and the ignition of fuel leaks, faster programmes to keep debris off airfields, and an ICAO study of automatic runway debris detection. BEA §4, pp. 178–180.
We remember the 113 people who died: the passengers and crew of Flight 4590 and four people on the ground in Gonesse.
For another investigation into an airliner opened by a hidden structural weakness, read what tore open Aloha Airlines Flight 243.
Primary sources and visual provenance
- BEA (Bureau d’Enquêtes et d’Analyses), final report on the accident of July 25, 2000, to Concorde F-BTSC, English translation (January 2002): synopsis and history of the flight, pp. 14–18; runway evidence, pp. 59–61; the DC-10 and strip examination, pp. 103–107; tank rupture studies, pp. 113–118; analysis, pp. 161–166; findings and probable causes, pp. 174–176; recommendations, pp. 177–180. Report PDF.
The article image and preview thumbnail are illustrative AI images from the episode’s reconstruction footage; the thumbnail adds editorial typography. Neither is a photograph or recording of Flight 4590. The episode uses AI narration and AI-generated illustrative scenes alongside simplified technical animations.
